Light source calibration method and system employed in source mask optimization

Abstract

Light source calibration methods and systems employed in source mask optimization are provided. The method includes: initializing a light source pattern and a mask pattern; using an SMO algorithm to iteratively optimize the light source pattern and the mask pattern; using a pre-established light source error correction model to correct the light source pattern after each iterative optimization, and updating the light source pattern after each iterative optimization with a corrected light source pattern in a current iteration process. The light source error correction model is established according to an input and output data set consisting of an input target light source pattern and an output actual light source pattern of a PIS. The method includes determining, according to an evaluation criterion or a condition of convergence of iteration of the SMO algorithm, whether the optimization meets a requirement.

The invention claimed is: 
     
       1. A light source calibration method employed in source mask optimization, comprising:
 (1) initializing a light source pattern and a mask pattern; 
 (2) using a source mask optimization (SMOG algorithm to iteratively optimize the light source pattern and the mask pattern; 
 (3) using a pre-established light source error correction model to correct the light source pattern after each iterative optimization, and updating the light source pattern after each iterative optimization with a corrected light source pattern in a current iteration process, wherein the light source error correction model is established according to an input and output data set consisting of one or more input target light source patterns and one or more actual light source patterns correspondingly output by a programmable illumination system (PIS); and 
 (4) determining, according to an evaluation criterion or an iteration convergence condition of the SMO algorithm, whether the optimization meets a requirement, and if so, ending the optimization, and outputting the final optimized light source pattern and mask pattern, or if not, returning to the step (2), wherein the evaluation criterion of the SMO algorithm involves inputting the iteratively optimized light source pattern and mask pattern to a lithographic imaging system model to perform lithographic imaging, performing calculation, and performing determination according to an image quality parameter. 
 
     
     
       2. The light source calibration method according to  claim 1 , wherein in the step (3), a method of establishing the light source error correction model comprises:
 (a) constructing a target light source data set of the one or more input target light source patterns, inputting each of the one or more input target light source patterns in the target light source data set to the PIS, and then acquiring each of the one or more actual light source patterns correspondingly output by the PIS; 
 (b) forming the input and output data set of the light source error correction model according to the one or more target light source patterns and the one or more actual light source patterns corresponding thereto; and 
 (c) using, according to the input and output data set, a nonlinear regression or neural network predictive model method to perform training and model assessment, and establishing the light source error correction model. 
 
     
     
       3. The light source calibration method according to  claim 2 , wherein in the step (a), Zernike polynomials or Legendre polynomials are used to generate a certain number of the one or more input target light source patterns to construct the target light source data set. 
     
     
       4. The light source calibration method according to  claim 1 , wherein the step (3) comprises:
 (3.1) setting an iteratively optimized light source pattern s k  to be an initial input of a correction process, wherein an initial light source pattern of the light source error correction model is s k, 0 =s k , k being the number of iterations in the SMO algorithm; 
 (3.2) setting a maximum number L of iterations in the correction process and setting a current iterator j to 0; 
 (3.3) updating the iterator j=j+1 in the correction process, and optimizing a light source pattern to s k,j  in the correction process by using an optimization method; and 
 (3.4) determining whether an iterator in the current correction process has reached the maximum number of iterations or determining whether the optimized light source pattern in the step (3.3) meets the condition |f(s k,j )−s k |<ε, ε being a set threshold, and if so, updating the light source pattern s k  after the current iteration using the SMO algorithm with the light source pattern f(s k, j ) corrected in the SMO iteration process, or if not, returning to step (3.3). 
 
     
     
       5. The light source calibration method according to  claim 4 , wherein in the step (3.3), the optimization method is a least square method, a regularization method, or a convex optimization method. 
     
     
       6. The light source calibration method according to  claim 1 , wherein when whether an iteration meets a requirement is determined according to the iteration convergence condition of the SMO algorithm in the step (4),
 the step (2) comprises:
 (2.1) setting a maximum number N of iterations in the SMO algorithm, and setting an iterator k to 0; 
 (2.2) updating the iterator k=k+1; and 
 (2.3) using the SMO algorithm to iteratively update the light source pattern and the mask pattern, so as to acquire a light source s k  and a mask m k  after the current k-th iterative optimization; and 
 
 the step (4) specifically comprises: determining whether the iterator k in the SMO algorithm is greater than the set maximum number N of iterations, and if so, ending the optimization, and outputting the final optimized light source pattern and mask pattern, or if not, returning to the step (2.2). 
 
     
     
       7. The light source calibration method according to  claim 1 , wherein in the step (4), the step of determining, according to the evaluation criterion of the SMO algorithm, whether an iteration meets a requirement specifically comprises: inputting the light source pattern and the mask pattern after the iteration to the lithographic imaging system model to perform the lithographic imaging; calculating an image edge placement error (EPE); determining whether the image edge placement error (EPE) is less than a set error; and if so, ending the optimization, and outputting the final optimized light source pattern and mask pattern, or if not, returning to the step (2). 
     
     
       8. The light source calibration method according to  claim 1 , wherein in the step (1), an SMO critical pattern selection method based on pattern clustering is used to select critical patterns from a layout to form an initialized mask pattern, or an SMO critical pattern selection method based on spectral analysis is used to select critical patterns from a layout to form an initialized mask pattern. 
     
     
       9. The light source calibration method according to  claim 1 , wherein in the step (1), the initialized light source pattern is set to a light source having a random pixel distribution or a light source having regularized parameters. 
     
     
       10. The light source calibration method according to  claim 1 , the one or more input target light source patterns being two or more input target light source patterns, and the one or more actual light source patterns being two or more actual light source patterns correspondingly output by a programmable illumination system (PIS). 
     
     
       11. A light source calibration system employed in source mask optimization, comprising:
 an initialization module, for initializing a light source pattern and a mask pattern, and transmitting same to an iteration module; 
 the iteration module, for using an SMO algorithm to iteratively update the light source pattern and the mask pattern, and transmitting same to a correction module; 
 the correction module, for using a pre-established light source error correction model to correct the light source pattern resulting from each iteration, and using the corrected light source pattern as an updated light source pattern in a current iteration, wherein the light source error correction model is established according to an input and output data set consisting of one or more input target light source patterns and one or more actual light source patterns correspondingly output by a PIS; and 
 a determination module, for determining, according to an evaluation criterion or an iteration convergence condition of the SMO algorithm, whether the iteration meets a requirement, and if so, ending the optimization, and outputting the final optimized light source pattern and mask pattern, or if not, performing iterative optimization on the light source pattern and the mask pattern again, wherein the evaluation criterion of the SMO algorithm involves inputting the iteratively optimized light source pattern and mask pattern to a lithographic imaging system model to perform lithographic imaging, performing calculation, and performing determination according to an image quality parameter.